Re: TAKE - Trade in panic() for BUG() in debug kernel

["Amit D. Chaudhary" <amitc@brocade.com>]

Just some thoughts.

Why not make this a configuration item. For stable usable systems, it is
more appropriate to reboot and do recovery. For all others like
qualification, testing the default approach below is better.

A similar case is the places where the inode is marked as bad,
make_bad_inode instead of a xfs_force_shutdown.

Regards
Amit


----- Original Message -----
From: Eric Sandeen <sandeen@sgi.com>
To: <undisclosed-recipients: ;>
Sent: Tuesday, May 21, 2002 3:34 PM
Subject: TAKE - Trade in panic() for BUG() in debug kernel


> There was a spot in here that would panic() the system,
> but that left the system useless for debugging; BUG() will
> let us get into kdb and see what's going on.
>
> Date:  Tue May 21 15:32:45 PDT 2002
> Workarea:
stout.americas.sgi.com:/localhome/src/sandeen/2.4.x-xfs/workarea-testing
>
> The following file(s) were checked into:
>   bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs
>
>
> Modid:  2.4.x-xfs:slinx:119800a
> linux/fs/xfs/xfs_da_btree.c - 1.126
> - Don't panic on the xfs_buftrace error, BUG() so we can still use kdb
>
>
>